mlsanityData to 7 Oct 2026

Does SAM dream of EIG? Characterizing Interactive Segmenter Performance using Expected Information Gain

arXiv 2404.16155 ↗First posted 24 Apr 2024cs.CV cs.IT cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

We introduce an assessment procedure for interactive segmentation models. Based on concepts from Bayesian Experimental Design, the procedure measures a model's understanding of point prompts and their correspondence with the desired segmentation mask. We show that Oracle Dice index measurements are insensitive or even misleading in measuring this property. We demonstrate the use of the proposed procedure on three interactive segmentation models and subsets of two large image segmentation datasets.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 15 of its 21 references that are in the map. First posted on arXiv 24 Apr 2024.

Builds on

15 of 21 references.

References · 15

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2023Segment Anythingcited by 8,158
  3. 2023Segment Anything Model for Medical Image Analysis: an Experimental Studycited by 291
  4. 2023Segment Anything Model (SAM) for Digital Pathology: Assess Zero-shot Segmentation on Whole Slide Imagingcited by 161
  5. 2023SAM-Med2Dcited by 123
  6. 2023Modern Bayesian Experimental Designcited by 121
  7. 2019Variational Bayesian Optimal Experimental Designcited by 116
  8. 2017On Nesting Monte Carlo Estimatorscited by 82
  9. 2023SAM.MD: Zero-shot medical image segmentation capabilities of the Segment Anything Modelcited by 61
  10. 2023Can SAM Segment Polyps?cited by 59
  11. 2023SA-Med2D-20M Dataset: Segment Anything in 2D Medical Imaging with 20 Million maskscited by 54
  12. 2023When SAM Meets Medical Images: An Investigation of Segment Anything Model (SAM) on Multi-phase Liver Tumor Segmentationcited by 52
3 more
  1. 2023Unleashing the Strengths of Unlabeled Data in Pan-cancer Abdominal Organ Quantification: the FLARE22 Challengecited by 50
  2. 2023Computer-Vision Benchmark Segment-Anything Model (SAM) in Medical Images: Accuracy in 12 Datasetscited by 45
  3. 2023The Segment Anything foundation model achieves favorable brain tumor autosegmentation accuracy on MRI to support radiotherapy treatment planningcited by 15

Built on by

Not cited yet.